arXiv Computer Vision research summaries for March 15, 2024.

Today's Research Themes (AI-Generated):

• ST-LDM: A novel universal framework for Text-grounded Object Generation (TOG) in complex real-world scenes using Swin-Transformer and latent diffusion models.

• CMR2D+T-SAM: A new approach for accurate cardiac MRI segmentation using SAM and spatio-temporal adaptation.

• QDMR: A comprehensive framework for real-world Computational Aberration Correction leveraging Unsupervised Domain Adaptation and a novel dataset.

• LOT: A novel framework for classifying point sets with spatial deformations using Linear Optimal Transport transform.

• Lifelong person re-identification (LReID) enhanced with backward-compatibility for practical application using cross-model compatibility loss.